Characteristics — Manufacture and Decoration — Classification — Potters' kilns

Potsherds are found on almost every Roman site and often in great abundance. It was an old opinion that the potter's wheel was a Roman introduction into this island, hence that 'thrown' pottery, unless imported, was no older than the Roman era; but it is now known that the natives used the wheel for two centuries or more before the conquest, and produced vessels of refined fabrique and artistic form. This Late-Celtic pottery, formerly classed as Roman, is found on Roman sites in the south of England, and there is little doubt that its manufacture survived the conquest unchanged. The term 'Roman pottery' is convenient and permissible, so long as it is understood to signify the ceramic products from whatever source, that were ordinarily used in Roman Britain.

This pottery, whether of home manufacture or imported, shows a marked advance in technique, and this was probably due to Roman influence; but this influence is less discernible in the forms and decoration. The work of the provincial potters has all the appearance of being substantially an indigenous development, and if it had a southern origin its prototypes must be sought in Italian and Greek forms before the advent of Rome as a world-power.

A notable exception, however, is the lustrous red pottery — the so-called 'Samian,' known on the Continent as 'terra sigillata' — which is found in considerable abundance in this country. It was not made here, and to the late Mr. C. Roach Smith stands the credit of first demonstrating that it was imported from the Continent. Subsequently, Dr. Dragendorff in Germany and M. D้chelette in France proved that it was manufactured in the valleys of the Loire and the Rhine from early in the 1st century to about the middle of the 3rd. From these centres it was dispersed throughout the empire, but especially in the western provinces and Italy. The fabric, however, was not indigenous to Gaul. Wares of the same kind had long been made in Italy, notably in and around Arretium, the modern Arezzo. It is significant that the manufacture declined in Italy in the same century that it appeared in Gaul, thus rendering it probable that the Italian potters migrated thither. This affords an explanation of the exotic character of this pottery on Gaulish soil; and it was the presence of this provincial redglaze which influenced the art of the local potters, whose imitations are known as 'pseudo-Samian.' The earlier examples of the ware resembled those of Italy, but gradually new forms arose and some of the older died out; the decoration, too, changed, but not to such a degree as to disguise its parentage.

It is almost impossible to convey by verbal description an adequate impression of the pottery of the era. This is best obtained by an inspection of a good collection, as that of Colchester, Guildhall, Reading, or York Museum. The following are some of its broader distinguishing features: There is an absence of white bodies which are so marked a feature in modern ceramic productions. The nearest approach is creamy-buff; but there is a preference for colours ranging from bright red, through tones of dusky maroons and browns to black, for the finer wares. There is an absence of painted subjects so characteristic of the Greek pottery, and of polychrome decoration so familiar to us. Painted work is comparatively rare, and is confined to simple stripes and scrolls, bold in effect, but often crudely executed. The prevailing decoration is in relief and generally displays considerable skill and artistic merit. Comparatively few have bright surfaces, and these, as a rule, are better described as lustrous or glossy, than as glazed. The material is earthenware: none has the hard and vitreous texture of our stoneware porcelain. The forms vary exceedingly. There are jugs, bowls and basins, shallow vessels of various shapes which only approximate to our saucers, plates, and dishes in their shallowness or their flatness, and others of shapes not represented in the ordinary vessels we use. On the other hand, we look in vain for forms resembling our tea- and coffee-pots, sauce-boats, and teacups. Less artistic than the Greek, the pottery nevertheless displays a gracefulness of curve not seen in the medieval, and not ordinarily in the modern. The vessels for the commonest purposes have an artistic feeling which contrasts with the severely utilitarian appearance of our culinary earthenware.

The methods of manufacture were simple. Although hand-made pottery was used — examples have been found at Silchester — it was exceptional. Broadly speaking, the wares were shaped on the wheel, but it is probable that the finest were finished on the lathe. The redglaze with raised figure and other subjects was, after leaving the thrower, pressed into moulds, and after removal, the feet were added, and lastly their interiors, the feet, and the external plain surfaces and beadings were finished on the wheel or the lathe; but moulding seems to have been rarely practised in this country.¹ The colour of the pottery depended largely on the clay used, but the potters were adepts at heightening or masking the natural colour. This was generally effected by a superficial wash or engobe, a process well known to the medieval and the modern potters. A vessel of dingy red clay, dipped, when in the 'green' state, in a thin mixture of fine pipe-clay and water, received a film which upon firing assumed a delicate cream colour. By the addition of yellow or red ochre, or of varying mixtures of the two, to the 'slip,' the resultant tint ranged from yellow-buff to salmon or pink. But for the finer wares there was a decided preference for a full red, and for various tones of deep warm browns and dusky maroons on the one hand, and for greys ending in black on the other. Some of these were certainly produced by the addition of mineral colouring agents to the engobe; but the darkest shades, and especially the greys and black, are due to the presence of carbon, sometimes as a superficial film, but more often it permeates the body as well. How the carbon was introduced is uncertain, and will be referred to later.

If by glaze is understood a translucent glass perceptibly distinct from the body although merging into it — as the glaze of modern porcelain — it is rarely seen on the pottery of the Roman era. Now and again fragments are found bearing a greenish-yellow glaze resembling that commonly seen on medieval wares, and apparently produced by the same method, that is, by dusting powdered galena (native sulphide of lead) over the clay pieces before firing. The glossiness of the redglaze more resembles that of the 18th-century saltglaze than a true glaze, that is, it appears to represent the surface of the pottery itself. Analyses have proved that the superficial portions of this ware are richer in soda than the interior, and it is not unlikely that the glossiness was the result of a chemical reaction between this alkali and the body-clay. Some of the finer black wares have a similar glossiness. Generally speaking, the dark brown and liver-coloured engobes have a faint waxy lustre, but not infrequently the finest dark wares have a bright metalloid surface, and even a slight iridescence. Occasionally vessels of fine texture have a smooth surface evidently produced by friction. These may be described as polished wares, and it is not unlikely that they were rendered bright by the application of wax.

Fig. 43. — A.	Fig. 44. — A.
Examples of Roman Redglaze (Terra Sigillata or 'Samian' Ware). (1/4)

The decoration of the pottery, however elaborate, is always in good taste: it never oversteps its proper province, or is so pronounced as to detract from the form. As already stated, the finest and most characteristic decoration is in relief. There were several methods by which it was produced, but first in importance is moulded work (Fig. 43, Nos. 1, 13, 17). The moulds, in which the decorated redglaze vessels were pressed, were of fine porous earthenware, unglazed, in order that much of the moisture of the clay pressed into them should be rapidly absorbed, and thus induce shrinkage and allow of the vessel being withdrawn. The mould was made on the wheel, and probably its interior was shaped by an iron 'profile'; then, while it was still moist, the decorative details were impressed from stamps of earthenware, metal, gypsum, and other materials. The bands of egg-and-tongue and other patterns were probably impressed from roulettes or wheel-like instruments, applied, in the case of the horizontal ones, while the mould was revolving. A comparatively small stock of these stamps admitted of innumerable combinations of decorative elements. Another method by which raised ornamentation was produced is occasionally seen on the finest redglaze (Fig. 44, Nos. 21, 34). The decorative details were made separately, each consisting of a piece of clay pressed into a metal intaglio and then applied to the surface of the vessel — a method in which Wedgwood among the moderns excelled; but it is usually combined with 'barbotine' decoration.

This barbotine or 'slip' decoration (Fig. 46, Nos. 5, 6, 8; Fig. 49, No. 5) is characteristic of the finer dark wares of Gaul and Britain, on which it is seen at its best. It was effected by the same or a similar process to that of the 17th-century potters, that is by trailing slip or thin clay upon the surface from a small vessel with a quill spout. The work had to be done rapidly, and its success depended upon an artistic instinct combined with unhesitating movement, both which qualities the Roman potters possessed in high degree. It was peculiarly adapted for scrolly designs, and the scrolls by the same movement of the hand could be made to terminate in disc-like or leaf-like expansions. These designs, simple as they are, are remarkably graceful and pleasing. But the clever decorators frequently essayed with equal success the task of delineating hounds chasing deer, and even human figures, as the gladiators engaged in combat on a large vase at Colchester (Fig. 41). A simple decoration consisting of lines of raised dots or studs arranged in oblong or lozenge-shaped patches (Fig. 45, Nos. 2, 7) is frequently met with, and it appears to have been produced by a comb-like tool alternately dipped in slip and applied to the side of the vessel. In barbotine work, the decoration was either of the same or of a different colour from that of the ground. In the latter, the trails were cream coloured, pale yellow, or red, which thus contrasted with the dark engobed surface of the vessel. In the former they were not necessarily of the same colour as the body, as in these cases the engobe was applied after the decoration.

FIG. 41. ROMAN POTTERY, COLCHESTER MUSEUM

Other varieties of raised decoration are occasionally seen. One may be described as finger-pressed work. In this the vessel, or some portion of it, appears to have been coated with a thick slip, which by the pressure of the finger was forced up into ridges. By this means various curvilinear diapers were obtained, of which the scale (Fig. 46, No. 2) and an irregular 'crocodile-skin' pattern are noteworthy. Vertical bands or 'pillars' of scale pattern were manipulated by the same process on strips of applied clay. In 'frilled' work the thrower gave the vessel one or more thin flange-like beads, and these were then waved by the alternate up and down pressure of the finger or some tool (Fig. 50, No. 7). In 'indented' work, the sides of the vessel were gently pressed in to produce a series of shallow flutings or other hollows, as in Fig. 42. 'Rough-cast' work was effected by coating the portions of the vessel to be so treated, with a thin slip, and then scattering over it coarsely powdered clay or pottery.

FIG. 42. ROMAN POTTERY, COLCHESTER MUSEUM

Sunk decoration may be conveniently divided into incised and impressed, but neither is a conspicuous feature of the pottery of the time. The common grey and black globular jars and dishes often exhibit a simple trellis made by a pointed tool, but so lightly so that the lines are less visible as grooves than as burnished strokes (Fig. 45, Nos. 4, 5, 9). An incised pattern is occasionally seen which consists of a band of concentric semicircles from which depend series of parallel lines stroked in with a comb-like tool — a pattern apparently suggested by the 'festoon-and-tassel' (Fig. 47, No. 2). Impressed work is a common feature of the 'pseudo-Samian' ware — a fine ware with a thin red engobe somewhat imitating, or perhaps it would be more correct to say, inspired by the redglaze. The stamps were apparently of wood, cut into the forms of simple rosettes, circles, notched segments of circles, and so forth. Both incised and impressed work was, however, more frequently accomplished on the wheel. The comb held against the revolving vessel gave rise to a band of parallel lines, and if moved up and down, to a wavy band of the same — a simple decoration often seen on the commonest wares. The hatched bands and surfaces frequent on all varieties of the pottery, and commonly-known as 'engine-turning,' were evidently impressed from notched wheels or roulettes (Fig. 46, Nos. 4, 12, 13). Sometimes a definite pattern — as the egg-and-tongue — was cut on the edge of the roulette, and bands of this character occasionally occur on the 'pseudo-Samian' referred to above. There is another and rare variety of sunk decoration, confined to redglaze, which may be called 'cut-work,' for it was certainly effected by gouges and V-shaped chisels. The cut-out portions normally take the form of vesica-shaped hollows, which are arranged to form stellate and other patterns. It is curious that the potters of the predecessor did not avail themselves of sgraffito decoration, that is the cutting through an engobe, in order to show a pattern in the colour of the body.

Painted decoration, as already stated, represents the least developed side of the potter's art of the period. It may be described as 'clay-painting,' and it differs from true barbotine, in the use of a thinner slip and its application with a brush. The patterns are similar, and it is not always easy to distinguish between trailed and painted work. Common pale buff wares, probably of Broseley clay, are often relieved with thin washes of red, but they rarely take the form of definite patterns. Marbled work may be conveniently referred to here. It is excessively rare, and was almost certainly imported. It appears to have been effected in the same manner as the marblings of the old Staffordshire potters, that is, by the partial blending of slips of several colours on the surface of the vessel.

The uses and ancient names of the different vessels are a difficult branch of inquiry. One thing, however, is clear: the vessels were essentially made for use. The distinction between 'useful' and 'ornamental' wares is modern, and came into prominence under Wedgwood and his contemporaries, who adopted classical models for their ornamental products. There is no evidence that the Gaulish and British potters copied antique Greek, Etruscan, or Oriental pottery to meet an antiquarian taste, or introduced novelties for purely display purposes. On the contrary, their shapes were those in vogue in their own day. Roman writers occasionally refer to various pottery vessels by name, and now and again mention their uses. There were vessels for the storage of wine and other comestibles — for culinary purposes — for the table; and others appropriate for religious rites and to hold the ashes of the dead. The large vessels for the transport and storage of wine, oils, figs, and other liquids and solids, were according to their shapes and sizes designated dolia, amphorae, cadi, etc. There were urnae for carrying water; urcei, ampullae, and lagenae, which corresponded with our jugs; poculi, or cups, of which there were various forms with special names, some borrowed from the Greeks; patinae, patellae, and catinae, probably dish and saucer-shaped vessels mostly for the battle; ollae and pelves for culinary and other household purposes; and other names of uncertain application. The attempts to identify the vessels to which these names applied are only partially successful; and so far as the pottery found on our sites is concerned, the task seems hopeless, for these Roman writers lived at different periods and referred mostly to the wares of Italy, whereas those of Gaul and Britain were of local origin or were modified by local influences.

No satisfactory classification of the pottery of the Roman era has yet been, or at present can be, devised. Any system that makes one feature to the exclusion of others, as the material or rather its colour, or the ornamentation, or even the form, important as this is, the basis of classification, is necessarily an artificial one. The ideal system would be one based upon the sources of manufacture, whether individual factories or regions where wares of distinctive character were made. But at present this is only possible in a limited degree. The Gaulish redglaze stands well defined from all other wares. Less definitely, the fine and mostly dark wares, characterized by the prevalency of the forms shown in Fig. 46 and of barbotine decoration, may be treated as another group; and as these were extensively made in the Nen valley in the vicinity of Castor, 'Castor' or 'Durobrivian' has almost come to be a general term for this kind of pottery wherever made. The red 'pseudo-Samian' ware represents another well-marked group and probably of Continental origin. We may similarly detach a few more groups, but there will remain a large irresolvable residue made anywhere where suitable clay abounded.

It would facilitate the study of the pottery if a definite terminology for the forms could be adopted; but this would be difficult to accomplish, for with few exceptions form merges into form in a tantalizing fashion. Dragendorff² did useful service by publishing the chief forms of the redglaze and giving a number of each, and his list has been extended by D้chelette³ and Walters,⁴ and no doubt will yet be added to as new forms are discovered. He arranged his forms in a systematic manner before giving them numbers. His first 14 examples are Italian, some of which are also provincial, the remaining 41 being Gaulish and German. In each series, they are arranged in the same order, beginning with dish-like vessels and ending with craters and tall vases; the sequence, however, ceases with the appended forms of D้chelette (23 in 1904), and with further additions the general numbering will become more arbitrary. It is obvious that if his system is extended to the pottery generally, the numbers would soon run into hundreds and it would be impossible to carry in the mind the forms they relate to.

As the pages of pottery figured in outline will give a better idea of the forms than written description; it is only necessary to supplement them with comments. The figures are from actual examples mostly in museums, and they include all the ordinary forms with a few of the rarer. As the interiors of the shallow vessels were exposed to view, consequently were carefully finished, one-half of their figures present their sections and internal profiles.

A. REDGLAZE ('TERRA SIGILLATA' OR 'SAMIAN' WARE)
(Figs. 43 and 44)

This ware as found in Britain is derived from three chief centres: La Graufesenque, the Condatomagus of the Ruteni, in the south of France; Lezoux in the Auvergne in central France; and Rheinzabern the ancient Tabernae Rhenanae, near Speyer on the Rhine,— but most of it is from the second. The manufacture of pottery at La Graufesenque was already old when the Romans appeared on the scene; but under their influence the Rutenian potters produced a fabric closely resembling the Arretine, between A.D. 50 and 100. At Lezoux, redglaze was made about as early, and it continued to about the middle of the 3rd century. The output must have been enormous — Dr. Plique unearthed, between 1879 and 1885, 188 furnaces, and recovered the names of about 3000 potters in the vicinity of the little town — and early in the 2nd century, the Arvernian products were exported throughout the western empire and even beyond. The Rhenish redglaze appeared about the beginning of the 2nd century and ceased about the middle of the following century. It is probable that the cessation of this and the preceding industry was due to the incursions of the Alemanni in A.D. 256-9. Redglaze was also made at St. Remy near Vichy, Banassac, and Montans in the south of France, and Westerndorf near Salzberg, but there is no evidence of exportation to Britain.

It will be noticed that the redglaze vessels figured are, with the exception of Nos. 32-3-4, bowls, basins, and various shallower forms which may be described as saucers, dishes, and platters. Most of these fall into two series, those, as Nos. 1 to 12 and 25, with an angled outline, that is with a more or less pronounced shoulder between the foot and the lip; and those with a curvilinear outline, as Nos. 14 to 24, and 26 to 28, the bowl No. 13 being of intermediate form.

Moulded decoration is almost confined in this country to the carinated, cylindrical and hemispherical bowls, Nos. 13, 1, and 17 [Dragendorff's Forms, 29, 30, and 37]. Of these, the first were the earliest, and disappeared about the end of the 1st century, the third surviving and holding the field in strong force for about a century or more, while the second, which are not common, probably disappeared in the 2d century. The general disposition of the ornamentation varies little. The lip is usually beaded. Then after an interval below it, is a narrow band of egg-and-tongue or some similar pattern; and this surmounts the decorated frieze. In the carinated bowls there is a second and less important frieze below the carination which itself is usually ornamented; and the earlier hemispherical bowls also have a second frieze. The decorative elements are extremely diversified, consisting of foliage, flowers, diapers, and figures of gods and goddesses, heroes, warriors, athletes, dancers, sphinxes, centaurs, mermaids, birds, beasts, fishes, etc.; and their combinations are equally diversified — on one frieze there may be a continuous scroll of foliage; on another a continuous hunting scene; on a third, figures in medallions or compartments, with intervening diapers, and so forth.

Moulded decoration also occurs on redglaze that imitates metallic vessels, especially patellae and small bowls with two flat ear-like handles, but examples are rarely found in this country, and if decorated, the decoration is confined to the handles.

Reliefs in applique, usually combined with barbotine, apparently survived moulding. They are confined to globular jars or ollae as Nos. 33 and 34, and other tall vessels as No. 32, that could not well have been moulded. The reliefs in applique are mostly mythological beings, personifications, and busts, the foliage and other subordinate details of the decorative scheme being largely in barbotine. Examples of this decoration are rare in this country; but a simple ornamentation of conventional ivy-leaves in barbotine — perhaps sometimes moulded — on the convex flanges of bowls and saucers of the forms of Nos. 18, 19, and 22, is common, and long preceded applique.

Of the plain vessels of the first series named above, Nos. 5 and 8 [Forms 33 and 31] are frequently found and were made to the close of the redglaze period. Nos. 3 and 4 [8] and 11 and 12 [16 and 15] are rare and probably early, and may be considered as the prototypes of the former. Nos. 2 [64] and 10, both in the Guildhall Museum, are rarer still. In the curvilinear series are two prevailing forms, the hemispherical and the campanulate, of which Nos. 16 and 24 may be taken as types. Of the former, small bowls with convex flanges, as Nos. 18 and 22 [35 and 38] are the most frequent, and the second had a long innings; the rest are rather scarce, especially No. 20 [81]. The campanulate form as in Nos. 23 [7], 24, and 27 [7], is also rare and undoubtedly early. The little basin, No. 28 [27], is freely found, and seems to have been made almost to the close of the provincial redglaze period. The curious mortarium with the lion-head spout, No. 21 [45], is a decidedly late form. The platters, Nos. 29 and 30 [22 and 17], are survivals of Italian prototypes, and No. 31, in the Colchester Museum, is most unusual. Nos. 32 [53] and 34 [72] are both uncommon, and have already been referred to.

In a general way, the earlier redglaze is thinner, harder, brighter and redder than the latter. Much of it is stamped with the makers' names, usually within the vessel on the bottom, but occasionally on the side externally. As a rule the name is in a sunk oblong label, but occasionally, especially in the German fabrics, this is in the form of a foot, a circle, or a half-moon. It is either in the nominative, with or without F or FE for fecit, or in the genitive, with O or OF for officina or M for manu. The names are mostly Gaulish, and the lettering often exhibits Gaulish peculiarities.

B. OLLAE OR JARS OF OTHER FABRICS THAN REDGLAZE
(Fig. 45)

Fig. 45. — B. Examples of Ollae or Jars in other Fabrics than Redglaze (1/4)

The examples figured chiefly differ in their lips, and broadly speaking the small beaded and cornice-like lips of Nos. 3, 4, and 5 are early, while the curved lips of Nos. 6, 8, and 10 occurred throughout the Romano-British period. No. 9 with its faint trellis pattern is a very common form in coarse black and grey wares, and was much used as a cooking-pot. The little 'poppy-head' vase, No. 2, as also No. 7, are in fine engobed ware, and No. 8, from London, has a bright plumbago-like surface. The cordoned bands of this and Nos. 6 and 11 are perhaps Late-Celtic legacies. No. 10 may be considered as a passage-form from the jar to the bowl.

C. OLLA-LIKE VESSELS, OR CUPS, OF OTHER FABRICS THAN REDGLAZE
(Fig. 46)

Fig. 46. — C. Examples of Olla-like Vessels, or Cups, in other Fabrics than Redglaze (1/4)

D. BOWLS AND BOWL-LIKE VESSELS OF OTHER FABRICS THAN REDGLAZE
(Fig. 47)

Fig. 47. — D. Examples of Bowls and Bowl-like Vessels in other Fabrics than Redglaze (1/4)

E. AND F. SHALLOW VESSELS (SAUCERS AND DISHES) AND AMPHORAE OF OTHER FABRICS THAN REDGLAZE
(Fig. 48)

Fig. 48. — E. and F. Shallow Vessels (Saucers and Dishes) and Amphorae in other Fabrics than Redglaze (1/4)

Fragments of large amphorae are constantly found on our Roman sites. These ponderous vessels of coarse buff or red clay were from 20 to 30 ins. in height, and No. 8 is a prevailing form, but they were often taller in proportion to their girth. The makers' names are often stamped on the handles, and indicate that they were derived, as a rule, from southern countries. Probably they owe their presence in this country to have been imported full of wine or oil. Small amphorae, as Nos. 6, 7, and 9, are much less common, and many, as the second, may be described as two-handled jars, the handles of these being often mere eyelets.

G. JUGS (AMPULLAE) OF OTHER FABRICS THAN REDGLAZE
(Fig. 49)

Fig. 49. — H. Examples of Jugs (Ampullae) in other Fabrics than Redglaze (1/4)

H. MISCELLANEOUS VESSELS OF OTHER FABRICS THAN REDGLAZE
(Fig. 50)

Fig. 50. — H. Miscellaneous Vessels in other Fabrics than Redglaze (1/4)

Nos. 1 and 2, both from Caerwent, are two types of handled beakers or cups, which are usually in common black ware, but are by no means plentiful. Gen. Pitt-Rivers found both types at Rushmore and Rotherley, some of his examples have been small eyelet handles.⁵ Nos. 3 and 8 (Guildhall and Silchester) belong to a large class of diminutive vases, usually in fine red or buff wares, which probably served a variety of purposes — to hold unguents, cosmetics, and the like, and as children's playthings, dice-boxes, etc. Nos. 4 and 5 are probably of Continental origin. The one, from a Colchester interment, is painted with light scrolls on a red engobe, and the other, a Bath example, has trailed scrolls covered with a blackish engobe. The tall vase, No. 6, from Silchester, is of fine red ware with light slip decoration. The 'frilled' tazza, No. 7, occurs in various 'coarse' wares and is not uncommon. The remaining illustrations are examples of covered vessels and indicate the usual shapes of the lids. No. 9 is of fine engobed pottery ornamented with 'engine-turning,' attributed to Castor, and decidedly rare; Nos. 10 and 11, two common grey cineraries from Colchester; and No. 12, a lid from Gellygaer.

I. EARTHENWARE MORTARS (MORTARIA, PELVES?) OF ALL FABRICS
(Fig. 51)

Fig. 51. — I. Mortars (Mortaria, Pelves?) in all Earthenware Fabrics (All 1/3)

The mortar was a highly specialized vessel, pan-shaped, with a concave interior studded with fragments of quartz or iron-slag pressed into the surface while soft, and with a strong overhanging rim and spout. It was used for triturating, mashing or mixing substances, especially foods, the hard fragments aiding the process and preserving the surface from abrasion. From the absence of pestles, it may be inferred that these were of wood. The rims vary considerably. Three types may be distinguished — the roll and bead (Fig. 51, A to D); the 'hammer-head' (H and I); and the vertical (J and K). The latter two appear to be derived from the first, which almost certainly was the earliest, and E, F, and G may be regarded as passage-forms between it and the second. Vertical rims are characteristic of the redglaze mortars (Fig. 44, No. 21) and its imitations, all the other forms being in ordinary red and buff wares.

The strong projecting rim led the writer to suggest many years ago,⁶ that the vessel was not ordinarily used resting on a table — its small bottom would render it unsteady in this position — but that it was inserted into a round hole large enough to receive the body and yet to allow the rim to rest upon its edge, as indicated in the section, Fig. 51. The fact that, while the internal surface and the rim are carefully finished, the exterior of the body is often left in a rough condition, goes far to confirm this conjecture.

POTTERS' KILNS

The remains of a considerable number of these kilns have been found in this country. They varied in shape, size, and construction, but all appear to have been on the same principle. They were subterranean structures with their summits level with the surface or slightly protruding. The simpler were circular, from 3 to 4 ft. or more in diameter, with a tunnel-like furnace on the floor-level. This, however, did not open directly into the oven which contained the vessels to be fired, but into a space below it with a perforated roof or diaphragm to allow the hot gases of the fire to ascend into the oven. It is evident that these small kilns were packed with the wares to be fired from the top, and this implies an opening large enough for the purpose. The opening also served as a chimney, but, unless restricted, it would be wasteful of heat. No doubt there was a simple contrivance for reducing it according to the requirements of the draught, or for closing it altogether. Some of the Continental kilns appear to have had a lateral opening for the introduction of the pottery and a small chimney or smoke-vent in the vaulted roof, and some of our larger examples may have had a similar arrangement.

The simpler kilns were lined with clay mixed with chaff or grass, and often with broken pottery or tiles, to mitigate the contraction under the action of fire. The perforated bottom or diaphragm was of denser clay, or of tiles specially made for the purpose — wedge-shaped, the wide ends resting on a set-off or ledge around the interior, and the points meeting in the centre and supported by a pier usually projecting from the back of the structure, but sometimes isolated. In the more elaborate kilns, the sides were constructed of curved bricks cemented with clay, and the roof of the furnace was often arched. Many kilns of this type have been found in the neighbourhood of Castor, and a group of four arranged crosswise and apparently fed from a common furnace-pit, near St. Paul's, in 1677.⁷ Two of simple construction similarly radiated from a common pit at Silchester.⁸

A larger kiln of different construction, found at Radlett, Herts,⁹ was somewhat oval in shape. In the centre was an oval pier, the space between it and the surrounding set-off forming a continuous flue, which was arched with broken bricks so arranged as to leave a number of gates. The floor above was "of clinkers and burnt clay laid loosely, over which was placed a thin layer of sand" — a mode of construction which would render it permeable to the heat of the furnace. Of five kilns near Lexden, Colchester,¹⁰ four were circular, and two of these were remarkable in having two furnaces each. The fifth was oblong, 5 ft. 4 ins. by 4 ft. 4 ins., and the under-structure was admirably arranged to support the perforated floor and at the same time to allow of the heat being well distributed under it, there being on each side of the flue three rectangular recesses.

Mr. Artis, in his account of the Castor kilns, gives some particulars as to the packing. It would seem that as each layer of vessels was placed, the packer's assistant followed with a layer of coarse hay or grass upon which he had laid small pellets of clay, each being covered with hay which was turned down over the edge before the next was deposited. Thus tier after tier was laid until the kiln was filled, the object of the pellets being to allow of the contents being removed without the risk of breaking the pottery. He was of opinion that the carbonaceous coloration of the black ware, referred to on page 156, was produced by smothering the kiln, that is, by closing its orifice, at a certain stage of the firing, thus confining the carbonaceous fumes from arising from the hay. It is probable that some such process contributed to the effect, but it is doubtful whether it alone would give the desired result.

The Author's Notes:

1. Portions of three different moulds for bowls (Form 37) were found at Pulborough, Sussex, in 1909, and several other examples have been found in this country.

2. Bonner Jahrbuecher, xcvi, xcix.

3. Les vases c้ramiques orn้s de la Gaule Romaine.

4. Catalogue of Roman Pottery, Brit. Mus.

5. Vol. i, pp103 and 113; vol. ii, p153.

6. Derbysh. Arch. Soc. xl, plate vii.

7. Illustrations of Roman London, p79.

8. Archaeologia, lxii, p328.

9. Proc. Soc. Ant. 2, xvii, p261.

10. Collectanea Antiqua, i, p1.